New chemistry for enhanced carbon capture: beyond ammonium carbamates

Carbon capture and sequestration is necessary to tackle one of the biggest problems facing society: global climate change resulting from anthropogenic carbon dioxide (CO2) emissions. Despite this pressing need, we still rely on century-old technology—aqueous amine scrubbers—to selectively remove CO2 from emission streams. Amine scrubbers are effective due to their exquisite chemoselectivity towards CO2 to form ammonium carbamates and (bi)carbonates, but suffer from several unavoidable limitations. In this perspective, we highlight the need for CO2 capture via new chemistry that goes beyond the traditional formation of ammonium carbamates. In particular, we demonstrate how ionic liquid and metal–organic framework sorbents can give rise to capture products that are not favourable for aqueous amines, including carbamic acids, carbamate–carbamic acid adducts, metal bicarbonates, alkyl carbonates, and carbonic acids. These new CO2 binding modes may offer advantages including higher sorption capacities and lower regeneration energies, though additional research is needed to fully explore their utility for practical applications. Overall, we outline the unique challenges and opportunities involved in engineering new CO2 capture chemistry into next-generation technologies.


Introduction
Rising atmospheric levels of carbon dioxide (CO 2 ) are the major contributor to global climate change, with annual emissions approaching 40 billion tonnes. 1 Nearly two-thirds of anthropogenic CO 2 emissions result from the combustion of fossil fuels, including coal and natural gas, for the global production of electricity. 1 In addition, CO 2 emissions are an inevitable byproduct of other industrial processes, including the production of cement, steel, and natural gas. 1 As a result, new technologies are needed to mitigate emissions from these industrial point sources during the gradual transition to cleaner fuels and building materials. One such proposed technology is carbon capture and sequestration or utilization, in which CO 2 is selectively removed from low-concentration emission streams (4-15% CO 2 ) prior to its permanent storage underground or conversion into more valuable products. 2 Building upon technology developed in the 1930s to purify crude natural gas, many have shown that aqueous amine scrubbers are currently the most technology-ready sorbents for CO 2 capture from ue emissions on large scale (Fig. 1a). 3 Aqueous amine scrubbers are effective because amines react selectively with CO 2 to produce carbamic acid intermediates, which rapidly react with a second equivalent of amine to produce ammonium carbamates; under aqueous conditions, ammonium carbamates and carbamic acids can further react with water to produce ammonium (bi)carbonates. 4 The captured CO 2 is then desorbed using heat and/or vacuum (temperature and/or vacuum swing), thereby regenerating free amines. Over the last ninety years, there has been signicant optimisation of the amine structure to maximize working capacities (i.e. the usable amount of CO 2 captured in an actual process) while minimising regeneration energies (i.e. the total energy input needed to heat the material and desorb CO 2 ). 5 However, aqueous amine scrubbers are still faced with several challenges, including: (1) low capacities (<3 mol CO 2 per kg solution or <15 wt%) due to dilution of the corrosive amines with water; 6 (2) poor oxidative stability of amines towards O 2 ; and (3) degradation in the presence of contaminants such as SO 2 , which reacts with amines similarly to CO 2 . 7 In addition, one aspect of aqueous amine scrubbers has remained largely constant: the products of their reaction with CO 2 . 8 This restriction generally leads to high regeneration energies ($2.4 MJ kg À1 CO 2 ) and CO 2 desorption temperatures (>100 C), greatly increasing the cost of carbon capture from ue emissions. [9][10][11] (ILs). Porous materials such as silicas, carbons, zeolites, metalorganic frameworks (MOFs), porous organic polymers (POPs), and covalent-organic frameworks (COFs), have the potential advantages of higher thermal stabilities and lower heat capacities compared to aqueous amine scrubbers. [16][17][18][19][20][21] Likewise, ILs are low-melting ionic salts that offer advantages over aqueous amines including non-volatility (preventing release into the atmosphere) and structural tunability. Although hydrophobic porous solids such as silicon-rich zeolites and carbons are capable of scrubbing CO 2 from high-concentration streams (e.g. crude biogas), 22 many of these materials cannot remove CO 2 from humid low-concentration streams such as ue gas emissions. 23 This limitation arises because CO 2 and water directly compete for the same physisorption sites in these sorbents. An additional general challenge for porous solid adsorbents that remains to be addressed is their poor thermal conductivity, which complicates adsorbent heating and cooling during adsorption/desorption cycling.
A powerful approach to overcome the poor selectivities of typical sorbents towards CO 2 under humid conditions is to leverage the favourable reactivity of aqueous amine scrubbers in the form of amine-functionalised sorbents (Fig. 1). 24,25 Beginning with the rst report of amine-functionalised silicas in 1992 (Fig. 1b), 12,24 a range of amine-functionalized solid adsorbents, including zeolites (Fig. 1d), 14 MOFs (Fig. 1e), 15,19 and carbons 26 have been prepared. Researchers have demonstrated that amine-functionalised porous solids possess the high CO 2 selectivities native to aqueous amines while generally evidencing improved thermal and chemical stabilities. For example, conning amines within a porous support largely eliminates oxidation pathways that are catalyzed by leached metal ions from the absorption columns. 7,27 Similarly, ionic liquids (ILs) can also be functionalised with amine groups to achieve high CO 2 capacities and selectivities without the need for dilution with water ( Fig. 1c). 28,29 Numerous in situ spectroscopic studies using solution-and solid-state nuclear magnetic resonance (SSNMR) and infrared (IR) spectroscopy combined with theoretical calculations suggest that in most cases aminefunctionalised materials produce similar sorption products as aqueous amine scrubbers, namely, ammonium carbamates under dry conditions 13,30,31 and, as conrmed recently, ammonium bicarbonates under humid conditions. 32 As such, the majority of these materials still require high temperatures (>120 C) to fully desorb CO 2 , resulting in high regeneration penalties. 33 In addition, amine-functionalised silicas suffer from oxidative degradation by distinct bimolecular pathways, 34 as well as the irreversible formation of ureas under dry conditions. 35 Overcoming these fundamental limitations is critical to enabling the widespread adoption of carbon capture technologies.

New CO 2 chemisorption pathways in solution and the solid state
An underexplored approach to overcome the fundamental limitations of amine-based materials is not to focus on the development of new materials, but on new chemisorptive pathways for selective carbon dioxide capture. For example, the formation of carbamic acids by CO 2 capture at amine sites is potentially desirable because it involves reaction with CO 2 at only a single amine site, increasing the CO 2 : amine sorption  13 (d) Post-synthetically amine-functionalised zeolites. 14 (e) Post-synthetically amine-functionalised metal-organic frameworks. 15 ratio to 1 : 1. 36 Indeed, unlocking 1 : 1 reaction stoichiometries in general should produce higher gravimetric and volumetric sorption capacities by enabling a higher density of reactive sites within a given volume. Additionally, mechanisms beyond ammonium carbamate formation have been shown to lead to lower CO 2 desorption temperatures in some cases (see below). Importantly, each combination of CO 2 partial pressure (P) and temperature (T) for a given separation (e.g. 400 ppm, 25 C for capture directly from the atmosphere) leads to an ideal differential free energy of sorption (ÀDG) for that separation (e.g. À19 kJ mol À1 for direct air capture), which is critical to maximising sorption capacities while minimising regeneration energies. 41 New chemisorption pathways should enable more dramatic tuning of the differential enthalpies (ÀDH) and entropies (ÀDS) of sorption to achieve these optimal values. Last, moving away from amines entirely could lead to adsorbents with improved oxidative stabilities, a recurring challenge associated with amine-based materials, although more work is required to characterize the oxidative stability of promising sorbents. 42 Here, we highlight examples of new CO 2 adsorption pathways beyond ammonium carbamates that may ultimately lead to enhanced CO 2 capture.
The unique, highly-charged environment within ILs makes them an ideal setting to unlock new CO 2 reactivity. For example, although carbamic acids are normally disfavoured outside of polar aprotic solvents (e.g. dimethyl sulphoxide), 36,43 Schneider, Brennecke, and coworkers found that installing amines onto the anions of amino acid-derived ILs favours CO 2 capture via the formation of carbamic acids stabilized by hydrogenbonding (Fig. 2a). 37 This change in mechanism doubled the molar absorption capacity of these ILs compared to those bearing amine-functionalized cations, which operate by the traditional ammonium carbamate mechanism. 13 In addition, the strong binding of CO 2 within a proline-derived IL (ÀDH abs ¼ 80 kJ mol À1 ) led to nearly complete saturation at low pressures of CO 2 (<0.1 bar at 25 C). Therefore, this switch in chemisorption products demonstrates that the local environment of an amine is a crucial design element for controlling its reactivity towards CO 2 . 44 Following these initial studies, even more unconventional CO 2 absorption pathways began to emerge in ILs. Building upon previous reports, 45 Li, Dai, and coworkers demonstrated that amines can be completely bypassed by capturing CO 2 in ILs bearing alkoxide or phenoxide anions and organic superbasederived cations, which reversibly capture CO 2 via alkylcarbonate formation (Fig. 2b). 38 Similar to carbamic acids ( Fig. 2a), this chemistry gives rise to a 1 : 1 reaction stoichiometry and thus higher gravimetric capacities (up to 20 wt%) compared to Proposed absorption mechanism by phenoxide and alkoxide ILs. 38 (c) Proposed mechanism for electrochemical CO 2 capture by 1,4-naphthoquinone. 39 (d) Proposed absorption mechanism for an IL with an aspartate dianion. 40 In all cases, the corresponding cations are omitted for clarity.
traditional IL sorbents (<10 wt%). Importantly, alkoxide-based ILs also possess low viscosities and rapid absorption kinetics (saturation in less than 5 minutes at room temperature), overcoming common challenges that plague traditional aminefunctionalized ILs. 38 Subsequently, Kim and coworkers demonstrated that similar reactivity at oxygen could be achieved in water-lean alcoholamines bearing sterically-hindered amines and that the resulting ammonium alkylcarbonates desorb CO 2 more readily than ammonium carbamates. 50 Another route to generate oxyanion nucleophiles for rapid CO 2 capture via carbonate formation is by the electrochemical reduction of quinones, as demonstrated by Hatton and others (Fig. 2c). 39 Promising results with electrochemically-reduced quinones have been observed in the presence of water and oxygen, although some loss in capacity was observed due to reoxidation of the nucleophile by oxygen. 51 This electrochemical approach has subsequently been expanded to other nucleophiles, such as reduced sulphides, suggesting it may be a general strategy to expand the scope of nucleophiles for CO 2 capture. 52 An advantage of this approach is that electrochemical regeneration of the quinone (electrochemical swing adsorption) leads to energy savings over traditional temperature or pressure swing processes.
Recent work has revealed that CO 2 capacities approaching a remarkable 2 : 1 reaction stoichiometry can be accessed in ILs, representing a four-fold increase compared to the traditional ammonium carbamate mechanism (Fig. 2d). 40 Specically, Wang and coworkers found that an ionic liquid with an aspartate dianion was able to reversibly bind 1.96 mol CO 2 per mol IL at 30 C and 1 atmosphere CO 2 , which was hypothesised to occur via two subsequent reactions at a single amine site to form both a carbamate (calculated DE abs ¼ À69 kJ mol À1 ) and a carbamic acid (calculated DE abs ¼ À54 kJ mol À1 ). 40 This proposed absorption pathway was supported by 13 C NMR measurements as well as density functional theory (DFT) calculations, with the latter ruling out reaction of CO 2 at the carboxylate groups as proposed for related ILs. 53 A similar 2 : 1 absorption mode was also evidenced in an earlier organic chemistry study. The observation of a triplet in solution 15 N NMR studies of selected primary amines in the presence of 13 CO 2 and a base conrmed the reaction of 2 CO 2 molecules with a single amine group at À30 C. 53 The high capacity offered by this absorption mode makes it a very attractive target for CO 2 capture applications.
Although ILs and water-lean solvents represent a unique platform for the discovery of new CO 2 capture products, they are not without their own challenges. For example, the absorption capacities of most ILs are relatively low (<20 wt%) compared to amine-functionalized solids. 13,28 In addition, the viscosities of ionic liquids are relatively high and tend to increase upon CO 2 adsorption (in some cases up to 200-fold), which represents a signicant process challenge. 28,54 While molecular engineering allows access to CO 2 -loaded ILs with viscosities as low as 650 mPa s, 55 these values are still signicantly higher than CO 2 -loaded 30% aqueous monoethanolamine solution (4 mPa s). 56 Last, the CO 2 /N 2 absorption selectivities, kinetics, desorption conditions, and long-term cycling stabilities of ILs remain poorly characterized in many cases. Addressing these challenges is critical to advancing the commercial viability of ILbased sorbents.
An emerging alternative approach is to engineer new CO 2 capture mechanisms within the controlled pore environments of crystalline porous materials, such as MOFs. The arrangement of functional groups in an ordered fashion within the pores of MOFs presents a potential opportunity for unlocking new CO 2 capture chemistry.
One of the earliest demonstrations of CO 2 chemisorption in MOF adsorbents involved CD-MOFs ( Fig. 3a; CD ¼ g-cyclodextrin). 46,57 These MOFs demonstrate strong adsorption of CO 2 at low partial pressures (<2 mbar), leading to excellent CO 2 /CH 4 selectivity (estimated to be >3000) in this regime. 46 Using SSNMR measurements, the authors proposed the formation of carbonic acids or alkylcarbonates; however, the exact chemisorption pathway in this material remains unclear. Nonetheless, the strong bonding of CO 2 in CD-MOF-2 (>1 mmol CO 2 per g MOF adsorbed at 10 mbar and 30 C) makes this a promising potential material for ue gas capture applications. Analysis of the thermodynamics of CO 2 chemisorption in this material by calorimetry revealed a moderate enthalpy of adsorption at intermediate loadings (ÀDH ads ¼ 65 kJ mol À1 ), enabling easier desorption of CO 2 from the strong-binding sites compared to amines. 58,59 However, the poor water stability of these MOFs necessitates the translation of this chemisorption mechanism to more stable materials for practical applications. 46 Carbamic acids have long been invoked as intermediates and products upon CO 2 capture in amine scrubbers, 4 aminefunctionalized silicas 30,60 and amine-functionalized MOFs, 61,62 as suggested by NMR and IR spectroscopies. For example, Ho and coworkers found that hydrazine-functionalized variants of the MOF Mg 2 (dobdc) (dobdc 4À ¼ 2,5-dioxido-1,4benzenedicarboxylate) exhibit incredibly strong and selective binding of CO 2 (3.89 mmol g À1 at 25 C and 0.4 mbar of CO 2 ), which they ascribe to highly favourable carbamic acid formation (ÀDH ads ¼ 90 kJ mol À1 ) within the framework pores. 62 However, until recently there remained little crystallographic evidence for this elusive adsorption product in the solid state. Long and coworkers identied variants of the MOF M 2 (dobpdc) (dobpdc 4À ¼ 4,4 0 -dioxidobiphenyl-3,3 0 -dicarboxylate) functionalised with the diamine 2,2-dimethyl-1,3-diaminopropane (dmpn) as promising adsorbents for post-combustion CO 2 capture owing to their exceptional hydrothermal and oxidative stability (Fig. 3b and c). 47,48 Exposure of single crystals of dmpn-Zn 2 (dobpdc) to 1 bar of CO 2 induced the formation of carbamic acid pairs bridging two adjacent amine sites in the framework, as conrmed by SCXRD and SSNMR (Fig. 3b). 47,48 In this structure, the normally disfavoured formation of carbamic acids is facilitated by well-dened hydrogen-bonding interactions, corroborated by the presence of strong 1 H (COOH) / 13 C correlations in 2-dimensional SSNMR experiments. Notably, carbamic acid pairs were actually predicted computationally in related frameworks before they were observed experimentally. 63 Building upon this work, the same group demonstrated that dmpn-Mg 2 (dobpdc) chemisorbs CO 2 by another distinct pathway: the formation of both ammonium carbamates and carbamic acids (Fig. 3c). 48 In-depth DFT calculations and 2dimensional SSNMR experiments support the formation of ammonium carbamate chains that interact with carbamic acids via hydrogen-bonding in this material. The advantage of this mechanism lies in its high enthalpy of adsorption (DH ads ¼ À74 kJ mol À1 ) coupled with a large entropic penalty (ÀDS ads ¼ 204 J mol À1 K À1 ), which reduces the temperature required to desorb CO 2 in a temperature-swing adsorption process to <100 C, potentially enabling adsorbent regeneration with lowgrade steam. 47 These thermodynamic parameters enable adsorbent regeneration with an estimated energy of 2.5 MJ kg À1 CO 2 , comparable to the best-in-class aqueous amine scrubbers such as Mitsubishi KS-1 (2.4 MJ kg À1 CO 2 ). 10,64 Therefore, this nding highlights the potential to overcome thermodynamic trade-offs of carbon capture processes by tuning the adsorption pathway. In addition, this adsorption mode leads to faster adsorption kinetics than ammonium carbamate formation in related materials and a high non-competitive CO 2 /N 2 selectivity (880) under the conditions relevant for CO 2 capture from coal ue emissions (150 mbar CO 2 , 750 mbar N 2 , 40 C). 65 Crystallographically confirmed formation of carbamic acid pairs in dmpn-Zn 2 (dobpdc) (dmpn ¼ 2,2-dimethyl-1,3-diaminopropane; dobpdc 4À ¼ 4,4 0 -dioxidobiphenyl-3,3 0 -dicarboxylate). 47 (c) Proposed formation of mixed carbamic acids and ammonium carbamates in dmpn-Mg 2 (dobpdc). 48 (d) Proposed formation of metal bicarbonates in Zn(ZnOH) 4 (bibta) 3 (bibta 2À ¼ 5,5 0 -bibenzotriazolate). 49 Gray, white, red, black, dark blue, sky blue, and green spheres correspond to carbon, hydrogen, oxygen, rubidium, nitrogen, zinc, and magnesium, respectively.
A further promising avenue to unlock new selective CO 2 capture reactivities in porous materials is to look to nature for inspiration. For example, carbonic anhydrase enzymes are responsible for the transport of CO 2 in the human body. Many members of this family operate by the reversible reaction of a zinc-hydroxide species (Zn-OH) with CO 2 to form a zincbound bicarbonate species (Zn-OCO 2 H). 66 In an early study, Zhang and coworkers demonstrated that high-valent monodentate metal hydroxides in the water-stable MOFs Mn II Mn III (OH)Cl 2 (bbta) and [Co II Co III (OH)Cl 2 (bbta)] (bbta 2À ¼ dihydrobenzo[1,2-d:4,5-d 0 ]bis([1,2,3]triazolate)) strongly bind CO 2 with high CO 2 /N 2 selectivity (>250), even under humid conditions. 67 These materials exhibit highly exothermic capture of CO 2 at low loadings (ÀDH ads > 100 kJ mol À1 ), necessitating regeneration using owing N 2 at 85 C (simulating a temperature-vacuum swing process). A regeneration energy of 2.7 MJ kg À1 CO 2 was calculated for [Co II Co III (OH)Cl 2 (bbta)], which is comparable with best-in-class aqueous amines. Closely mimicking the mechanism of carbonic anhydrase enzymes, Wade and coworkers subsequently found that Zn-OH centers in the air-stable MOF Zn(ZnOH) 4 (bibta) 3 (bibta 2À ¼ 5,5 0 -bibenzotriazolate) strongly bind CO 2 to form metal-bound bicarbonates with adsorption capacities of 2.2 mmol g À1 at 27 C and 0.4 mbar of CO 2 , suitable for direct air capture (Fig. 3d). 49 Interestingly, DFT calculations suggest that these metal-bound bicarbonates hydrogen-bond with adjacent Zn-OH centers, stabilizing the adsorption product (ÀDH ads ¼ 71 kJ mol À1 ) and leading to steep uptake of CO 2 at low pressures. Dincȃ and coworkers demonstrated a similar bioinspired approach to CO 2 capture in (Zn 5 (OH) 4 (btdd) 3 ) (btdd ¼ bis(1,2,3-triazolo[4,5b],[4 0 ,5 0 -i])dibenzo[1,4]dioxin), a hydroxide-substituted variant of the MOF MFU-4l. 68 This framework was found to exhibit stronger CO 2 binding (ÀDH ads ¼ 81 kJ mol À1 ) compared to Zn(ZnOH) 4 (bibta) 3 , albeit with a lower adsorption capacity at low pressures (0.9 mmol g À1 at 25 C and 23 mbar of CO 2 ). In all of these studies, the formation of metal-bound bicarbonates was validated primarily by in situ IR spectroscopy and DFT calculations. Subsequent work by Wade and coworkers has highlighted the importance of metal identity on CO 2 adsorption in M-OH MOFs, unveiling a potential handle for tuning the thermodynamics of chemisorption. 69,70 In a similar vein, Wang and Lackner have found that hydroxide-functionalised ionexchange membranes are promising for energy-efficient moisture swing sorption processes, demonstrating that ammonium cations can be used as an alternative to metal ions to prepare hydroxide-rich materials. 71 The capture of CO 2 with oxygenbased nucleophiles in both adsorbents ( Fig. 3a and d) and solution ( Fig. 2b and c) represents a promising solution to overcome the inherent limitations of amine-functionalised materials; however, more work is required to map out the stability of these materials and their performance under realistic conditions.
The vast majority of CO 2 capture processes discussed above operate via the addition of nitrogen-or oxygen-based nucleophiles to CO 2 . Recently, the range of nucleophiles that can reversibly react with CO 2 has been expanded to include electrochemically-generated sulphides, 52 frustrated Lewis pairs, 72 and N-heterocyclic carbenes, 73 among others. 74 In addition, electrochemistry has emerged as a powerful tool to expand the scope of CO 2 capture processes. For example, Hamelers and coworkers have shown that capacitive charging and migration of bicarbonate/carbonate ions through ion exchange membranes can drive a CO 2 capture process with a low energy requirement of 40 kJ mol À1 CO 2 captured. 75 Similarly, Landskron and coworkers have developed a related process in which a supercapacitor device reversibly adsorbs CO 2 (<0.1 mmol g À1 ), although the exact adsorption mechanism remains unclear. 76 Finally, electrochemically driven pH swings are also being investigated as a new energy-efficient CO 2 capture strategy. 77 These recent directions represent an exciting opportunity to unlock new carbon capture chemistry.
Opportunities and challenges for nextgeneration CO 2 capture The foregoing examples highlight the unique opportunities offered by new CO 2 capture pathways. Potential advantages of new sorption modes include lower regeneration energies, higher working capacities, and access to a wider range of sorption enthalpies compared to traditional ammonium carbamate formation. Despite this promise, there remains a great need for additional research to assess the application of novel sorption pathways in industrial processes. Key materials challenges (Fig. 4) that remain critically underappreciated include: (i) rapid sorption kinetics, (ii) large working capacities under realistic mixed-gas conditions, (iii) sufficient material stability to survive long-term exposure to reactive contaminants in target gas streams, such as water, oxygen, sulphur dioxide, and hydrogen sulphide, among others depending on the process, 23,24 (iv) adsorbent development and structuring to overcome issues with low thermal conductivities and heat management during highly exothermic sorption processes, and (v) sustainable and scalable materials synthesis. As an example, the hydroxide-based MOF [Co II Co III (OH)Cl 2 (bbta)] has a promising working capacity and regeneration energy for a ue gas capture process, but its stability in the presence of oxygen and other contaminants remains unknown. 67 Moreover the very large heat of CO 2 adsorption at low loadings (ÀDH ads > 100 kJ mol À1 ) suggests that heat management will be an important challenge for this material. Promising strategies to aid heat management include the development of structured hollow bre adsorbents and the design of carbon-based materials that have inherently larger thermal conductivities. 78 Exploration of entirely new CO 2 capture pathways beyond those described in this perspective should also lead to further advances. Recent studies have highlighted the promise of largescale computational screenings in the search for new CO 2 capture materials. For example, Smit and coworkers recently reported the screening of over 300 000 theoretical MOFs and identied classes of physisorption sites, termed "adsorbophores", that endow high CO 2 selectivities to frameworks. 79 The guided synthesis of optimised materials for operation under humid conditions was achieved by selecting candidates with hydrophobic adsorbophores to maximize the adsorption of CO 2 under humid conditions. However, the capacities and CO 2 /N 2 selectivities reported for best-in-class physisorbents are typically lower than those reported for chemisorptive materials. Similar computational screens to predict chemisorption-for example, using a higher level of theory to account for bondbreaking andforming processes-remain rare but have the potential to be transformative. 63,80 Similarly, calculations that can predict chemisorption thermodynamics under realistic mixed gas conditions should lead to promising materials for real-world applications. 79 Due to the complex processes inherent to chemisorption, an additional challenge for computational analyses is to predict transition states relevant to sorption kinetics. A promising strategy to address these computational challenges may be to use machine learning to guide the search for new chemisorbent materials. 81 In order to elucidate and ultimately build upon new CO 2 capture chemistry, advanced characterisation methods are also needed. These methods serve to both validate and discover new chemisorption products when unexpected sorption properties arise. Recent years have seen signicant advances in the characterisation of CO 2 capture pathways through in situ spectroscopic and X-ray diffraction experiments. 30,32,47,48 These experiments must now be adapted to study conditions that more closely mimic envisaged industrial applications and, in particular, must address mixed gas conditions rather than pure CO 2 . 32,48 Furthermore, experiments must not be restricted to studying static/equilibrium conditions and should probe the dynamic conditions associated with practical sorption processes.

Conclusions
New CO 2 sorption pathways such as those recently uncovered in appropriately-functionalised ILs and MOFs may offer improved performance for CO 2 capture compared to traditional sorbents, including higher capacities and lower regeneration costs. Many of these binding modes do not readily occur in aqueous solution and instead arise from the unique opportunity to precisely install chemical functional groups with a controlled spatial arrangement and carefully tuned local environment. For many of these prospective sorbents, more work is needed to assess their sorption kinetics, selectivities, stabilities, and thermal conductivities. This mechanistically-focused line of sorbent discovery is still in its infancy, and a new generation of computational, analytical, and synthetic chemistry is needed to design transformative materialsand sorption mechanismsfor reducing anthropogenic CO 2 emissions.

Conflicts of interest
The authors declare the following competing interest: P. J. M. is listed as an inventor on several patents related to the preparation of metal-organic frameworks for CO 2 capture.